Usage experience and improvements

[rabbitii]

After some time i used this project i feel that it need some upgrade.

I need:

1. possibility to export config file of single drive from the main config for removing drive from the pool. It should check shared parts of plots and asking to delete shared plot from pool or relocate the shard to another drive if possible.
2. option to import exported config with mapping to the current drive link, because name of the drive may be changed when transferring to another machine.
3. receive list of files stored on device by device id with indication if whole file stored on this drive or only part of it.
4. option to create subfolders and assign files to it. Many forks harvesters can't work with more than 1000 plots at once. So, i need option to divide whole pool to parts.
5. changing in fulfilment strategy. Now the tool checking only the free space on the drives and writing plots to the drive with the biggest free space. When transferring new files in parallel it creates bottleneck and slowdown the pool because all imports may go to the same drive when idle drives with compatible free space are present. It would be a great upgrade if the tool will check also if there are additional idle drives with free space for new plot are available for writing and write the file to that drive. If all the drives currently busy add possibility to create queue and starting add new plot only when idle drive/s presents.
6. create additional command line switch for do not store plots to many shards and wait for disk with compatible free space (if there are current plot/s import/s presents) or exit.
7. do not publish plots from the absent drive/s in the mounted directory (startup check).

I understand the point of free time and donations, so if you want payment for those features, post price list per feature, please.

Thanks!

[benengele]

It's possible to convert the binary configuration file from/to JSON. See https://google.github.io/flatbuffers/flatbuffers_guide_tutorial.html and https://stackoverflow.com/questions/49935556/how-to-convert-flatbuffer-data-file-to-json You could probably use JSON import / export and a text editor to satisfy 1-3.

You could use Docker to run multiple plotfs instances for different folders: https://github.com/benengele/PlotFS-docker

[johnathon-b]

I've got Powershell scripts to export to csv by plot_id and device_id. It can be expanded further to capture shards and be pushed into any data format you want.

As @benengele mentioned, the bin file needs to be converted with flatbuffers to JSON.

A feature request that could open the door to solutions to your list would be the ability to export the bin file to JSON within the plotfs cli commands. Logic can be wrapped around that to give us most of what we need to manage plotfs.